Arq Backup and Amazon Glacier

After some issues with TimeMachine, I recently switched up my backups from TimeMachine/Crashplan+ to to CarbonCopyCloner/Arq Backup.

My choice to move to Arq Backup was fueled by the fact that while Crashplan+ Family Unlimited costs $149.99/year, Amazon offers storage on S3 for $0.130 per GB/month and on Glacier for $0.01 per GB/month. With the ~100 GB (inclusive of incremental backups) that I want backed up in the cloud, it will only cost me about $19.20/year (~95% of my data on Glacier) to store my backups with Amazon. Adding in the few dollars it will cost me in data transfer, PUT, and GET request charges from S3/Glacier, my yearly cost certainly won’t exceed $25-30. I justify storing most of my data in Glacier because I make regular backups of my home directory to my NAS using Carbon Copy Cloner. Every few months, I also make a full clone to an external drive. The data I backup to S3 is whatever I keep in Dropbox — any critical files and files I am currently working on.

What’s the catch? Amazon Glacier charges you a retrieval fee of $0.01/GB if you need to retrieve more than 5% of your average monthly storage, Also, if you delete data within 90 days of upload, you are charged a prorated $0.03/GB early deletion fee.

Within Arq, a great feature is the ability to limit your storage on S3 by either a monthly dollar or GB amount. When you hit your max, Arq automatically prunes the oldest incremental backups (leaving at least one for each backup folder). While many people complain about the lack of this feature for Glacier, it makes sense that it is not included due to the early deletion fee. After a great conversation on ADN, I learned that many were having trouble manually deleting their Arq Glacier backups. Since it looked like the feature was there, I decided to contact the developer to get the question answered.

Stefan responded:

If you remove a folder from Arq that’s being backed up to Glacier, Arq puts the folder in its “trash”.
If you then open Arq’s trash, select that folder, and click “Delete Permanently”, Arq will delete all the Glacier archives and attempt to delete the Glacier vault. The vault delete will fail because Amazon has to update its “inventory”, which it does once/day. The next day, browse under “Other Backup Sets” in Arq, find that vault, select it and click “Delete” to delete it.

It seems that the confusion stems from the fact that Amazon only inventories Glacier vaults once a day. Once you delete the backup folder from Arq, it gets disconnected from your Arq set and once the inventory is updated (sometime within the next day), you can then delete the backup on Glacier.